Process variation may lead to lower yields in semiconductor manufacture. One conventional approach for detecting process variation includes the use of test patterns. One conventional test pattern includes a resistor formed on a substrate. Active regions are formed in the substrate material by doping, and one or more of the active regions can be used as a resistor. To form the resistor, metal contacts are placed along the length dimension of an active region. Resistance is then measured between the contacts.
It is generally assumed that process variation, which may result in feature dimensional variation, doping variation, and the like, may affect the resistance of the test pattern. Thus, process variation may be detected by discerning a difference in measured resistance versus another resistance value. It is also generally assumed that process variation affecting a test pattern may affect one or more other portions (or the entirety) of the wafer.
However, some conventional test patterns are not sensitive enough to provide reliable indications of some process variation. What is needed is a test pattern capable of providing more robust results.